Prediction Of Stroking Characteristics Of Elite Rowers From Anthropometric Variables


Physique is an important element of performance in rowing and determines, among other things, adaptability to a long lasting training programme [18].

Some evidence indicates that anthropometric characteristics may have some influence on rowing performance. Champion rowers tend to be taller [5] and heavier [16] than their less successful counterparts. In addition, they typically have lower subcutaneous skinfold scores [5]. Many have long limbs, not only in absolute terms but also in proportion to their standing height [5, 14]. Anthropometric studies of lightweight rowers competing at the 1985 world championships showed that they, too, tended to be tall, lean, and long limbed when compared with a reference population of university students [13]. It seems clear that the assessment of rowers should include various anthropometric characteristics.

Anthropometric data from elite adult rowers underlines the importance of body mass and body size. Nevertheless, while there is a lot of interest in the relationship between anthropometry and rowing performance, little attention has been focused on the prediction of the stroke rate and stroke lengths from the anthropometric profile of the rower. There are no data describing the relationship between limb and limb segment lengths and rowing performance. Moreover, it is not known how arm and leg lengths impact on stroke rate and force production. The purpose of the present study was to investigate how the anthropometric characteristics, and more specifically the limb lengths of the rowers, impact on the self selected stroke rate and stroke lengths during rowing at different power outputs. The results of the present research will improve our understanding of the importance of rowers’ anthropometric profiles on rowing performance, and, more importantly, will provide insight into its impact on self selected SL and SR. Such information has implications for crew selection.

Material and Methods

Eight, male competitive rowers (age 17.0 ± 0.5 yr, height 180.8± 3.1 cm, weight 76.5 ± 3.3 kg), volunteered for the study. The subjects were randomly selected from the male participants at the Greek Open Junior National Championships. At the time of the investigation, the rowers were in their competitive training phase and had no injuries that would affect rowing performance.

Each rower read and signed a written informed consent prior to participation and also their current health status was determined prior to testing.


A Concept II rowing ergometer (Concept II, Morrisville, Vermont) was used in this study since it accurately simulates many phases of the rowing stroke [21]. Subjects were video taped by Panasonic RX5 videocamera (Model NV-RX5EG, Matsushita Electric Industrial Co, Japan). The video camera has digital AI auto focus wide-angle video zoom lens. Heart rate was registered by means of Polar Vantage NVTM (Polar Electro OY, Finland). The basic components of the Polar Vantage are a watch and a chest transmitter. The athletes were weighed to the nearest 0.1kg on an electronic scale (Seca Alpha Model 770). The athletes’ heights were measured by a girth tape fixed to a wall with range of measurement of 600mm to 2100mm. Limb lengths were obtained using a girth tape.

Procedure and Measurements

Testing Procedure

Testing took place on 2 separate days. Physical performance testing was performed on day one while all anthropometric data were collected on day two. Orientation to the test procedures was conducted for all subjects.

Before starting the test, the chest belt of the polar heart rate monitor with the transmitter and the watch were placed on the athlete. The rower warmed up on the rowing ergometer for approximately 3 minutes at a self-selected intensity. After a short rest of 30 s, the rower commenced 3 maximal stroke rates followed by static stretching for 2 min. The test started when the heart rate of the subject was approximately 120 to 125 bpm.

All the tests were performed on the same Concept II rowing machine. The resistance level on the Concept II rowing machine flywheel was set to 10 for all testing. During the test the temperature was 28o C. The present investigation utilized a series of seven trials; each subject completed all the trials in a single session. During each trial the time remaining to the completion of the test was continually given to the rowers. The researcher and the coach provided strong verbal encouragement to all rowers during all trials.

The ergometer screen of the Concept II was programmed to display the 30- s time and the power (Watts), the drag factor and the stroke rate. The subject was asked to row at maximal stroke rate and intensity (100%) for 30 s. Immediately after the completion of the trial, the average power and average stroke rate and heart rate of the trial were recorded. Power outputs equaling to 60%, 65%, 70%, 75%, 80%, 90% of the maximum power (Watts) were then calculated. Subjects then rowed at the target power outputs using their own preferred stroke rate and stroke length. The intensities were randomized. When the subject reached the desired power, rowing kinematics was recorded by video camera for 5 – 6 rowing strokes. The signal to start the video camera was given from an assistant who had visual contact with the monitor of the rowing ergometer. Each subject was filmed for 5 – 6 rowing strokes while first rowing maximally, then at 90, 80, 75, 70, 65 and 60% of peak power. Heart rate and stroke rate were recorded. The same procedure was performed for all power outputs. The interval between the trials was approximately 2 minutes. The next trial commenced once the subject’s heart rate fell to 120 beats per minute.

Motion Analysis

A video camera placed 6 m from the subject and perpendicular to the line of rowing stroke was set to record the trial.

Stroke length, stroke rate and angular displacement and velocity of knee, hip and elbow joints were calculated, the arm moved largely in the sagittal plane and was observed not to vary noticeably between rowing intensities. For the calculation of these variables the Biokin software (Biokin V 4.0, Greece) was used after the data were filtered using a 4th order, zero-lag filter with a 6 Hz cut- off frequency.

Anthropometric Measurement

An anthropometric profile of the participants was taken. The profile included the weight, the height and limb lengths for each individual and all these components were measured according to the methods of Claessens et al. [3]. Specifically, the following variables were measured: weight, height, sitting height, arm length, forearm length, hand length, thigh length and tibia length.

Statistical Analysis

All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS 10). Descriptive analysis was used for all variables. Pearson’s Product moment correlation coefficient was used to reduce some anthropometric variables. Multiple analyses of variance (MANOVA) and Post Hoc Test (Tukey’s) were used to determine the differences on stroke length (SL) and stroke rate (SR) at different intensities. Regression analysis (stepwise methods) was used to predict stroke rate and stroke lengths at different intensities from the anthropometric characteristics. The Alpha level was placed at 0.05 for all tests.


Descriptive statistics of the physical characteristics of the rowers is presented in Table 1. The age, height and weight of the rowers were 17.0 ±0.5 yr, 180.8 ± 3.1 cm and 76.5 ± 3.3 kg, respectively. The means and standard deviations for all segmental lengths measures taken and the range are shown also in Table 1.


For these anthropometric variables the Pearson product moment correlation coefficient was used to determine the linear relationship between variables. All correlations between segment lengths were conducted in order to reduce the number of variables used in analyses. Consequently, the highest correlation was revealed between sitting height and standing height (r = 0.8, p = 0.01). A significant correlation was found between thigh and tibia length (r = 0.7, p = 0.04). No significant correlation was found between arm and forearm length (r = 0.5, p = 0.1).

There were no significant differences (p = 0.7) in stroke length between the trials at different intensities despite a trend toward increasing stroke length at lower intensities. This may in part be due to the large inter-individual variation in stroke lengths. Figure 1 shows the stroke lengths selected at each rowing intensity.

There were significant differences (p < 0.001) in stroke rate between the trials at different intensities. The multiple comparisons showed that there were no significant differences in stroke rate between maximum trial and 90% of peak power, between 80% of peak power and 75% and 70% of the peak intensity. Moreover, no differences in stroke rate were found between 65% and 60% of peak power. Figure 2 shows the stroke rates selected at each rowing intensity.

Stepwise regression analysis was performed to predict the stroke length and stroke rate at different stroke intensities. For the maximum intensity the best predictor for the stroke length was arm length (adjusted R2 = 0.98; see Table2). For all the other intensities arm length was again the best predictor.

Furthermore, the results of the regression analysis indicated that at the maximum intensity the best predictor for the stroke rate was thigh length (adjusted R2 = 0.98). For the intensities of 90%, 70% and 60% of the peak power the best predictor was sitting height (adjusted R2 = 0.98, 0.97 and 0.98, respectively). Arm length was found to be the best predictor for the intensities of 80%, 75% and 65% with adjusted R2 of 0.98, 0.98 and 0.98, respectively.


Practical Application

The model and the statistical treatment of the data gathered on the anthropometric variables included in the model supported the notion that selected segment lengths are important determinants of rowing performance.

More specifically, the results indicated that arm length, thigh length and sitting height – variables that are genetically determined – are related to the stroking characteristics at different intensities. The findings of the present study suggest that coaches could put rowers of the same anthropometric characteristics in the same boat and create a more successful rowing crew. It is expected that rowers of the same sitting height, arm and thigh length will have the same stroking characteristics (SL and SR). However, more research is required before making broader conclusions.